Fluorous-based small-molecule microarrays for the discovery of histone deacetylase inhibitors

Article abstract of DOI:10.1002/anie.200703198

(Figure Presented) Noncovalent immobilization is an attractive method for identifying inhibitors of histone deacetylases (HDACs). Fluorous-based small-molecule microarrays were validated as an effective method. Three enzymes and three assays (microarray, biochemical activity, and surface plasmon resonance) were used to identify inhibitors of HDACs and to compare them.

Full text of DOI:10.1002/anie.200703198

Communications
DOI: 10.1002/anie.200703198
Microarrays
Fluorous-Based Small-Molecule Microarrays for the Discovery of
Histone Deacetylase Inhibitors**
Arturo J. Vegas, James E. Bradner, Weiping Tang, Olivia M. McPherson, Edward F. Greenberg,
Angela N. Koehler, and Stuart L. Schreiber*
Reversible acetylation plays a key role in defining chromatin
states and in regulating transcription from genomic DNA
differentially across distinct tissues.[1–3] Histone deacetylases
(HDACs) function in this process by catalyzing the hydrolysis
of N-acetyl groups on lysine residues found in the N-terminal
tails of histone proteins.[4] This process mediates cell differ-
entiation, correlates with epigenetic inheritance, and is
deregulated in human disease, among others.[1–3,5]
Identifying novel HDAC inhibitors is an increasingly
active area of research.[6–10] Trapoxin, which is a naturally
occurring HDAC inhibitor, was instrumental in the original
discovery of HDAC1.[4] Suberoylaniline hydroxamic acid
(SAHA/vorinostat), which inhibits multiple members of the
HDAC family of enzymes, has been approved recently for the
treatment of cutaneous T-cell lymphoma.[11–13] Tubacin, which
is the first homologue-selective inhibitor (having selectivity
for HDAC6), has illuminated the function of HDAC6 and
tubulin acetylation.[14–19]
reported uses of SMMs to identify new HDAC inhibitors,
including ones having selectivity for specific members of the
HDAC family. Traditional SMMs use various chemistries to
attach compounds covalently.[20–25] Many of these approaches
either take advantage of latent functionalities that result in
heterogeneous molecular display on the surface, or require
synthetic modification of compounds to obtain homogeneous
display. Fluorous tags are versatile tagging groups for
chemical-library synthesis and can facilitate noncovalent
immobilization on fluorinated glass surfaces.[26–31,37] A pre-
vious report from Pohl and co-workers demonstrated fluorous
microarrays as a powerful screening tool for carbohydrate-
binding proteins.[26] Herein, we demonstrate that fluorous-
based SMMs enable screening for HDAC inhibitors by
allowing controlled molecular display of inhibitory function-
ality, low uniform background signals, and excellent signal-to-
noise ratios.
SMMs were evaluated as a tool for identifying HDAC
Biochemical, enzyme-activity assays involving fluorescent
readouts are frequently used to identify new HDAC inhib-
itors. However, this approach requires expensive reagents and
equipment, and can be difficult to perform in a high-
throughput manner. Small-molecule microarrays (SMMs)
provide an attractive alternative for high-throughput iden-
tification of HDAC inhibitors. Currently, there are no
binders or inhibitors by using a three-part validation
(Figure 1). Quantitative fluorescence data were collected
from probed arrays and used to generate a list of positives.
Non-fluorous tagged equivalents of the compounds were then
tested in a fluorescence-based biochemical activity assay with
the same set of enzymes to determine enzymatic inhibition.
Furthermore, thermodynamic and kinetic binding data were
collected for non-fluorous-tagged compounds binding to one
of the HDACs by using surface plasmon resonance (SPR)
[*] A. J. Vegas, Dr. J. E. Bradner, O. M. McPherson, E. F. Greenberg,
Dr. A. N. Koehler, Prof. Dr. S. L. Schreiber
Howard HughesMedical Institute
Chemistry and Chemical Biology
Harvard University
Broad Institute of Harvard and MIT
7 Cambridge Center
Cambridge, MA 02142 (USA)
Fax: (+1)617-324-9601
E-mail: stuart_schreiber@harvard.edu
Homepage: http://www.broad.harvard.edu/chembio
Prof. Dr. W. Tang
School of Pharmacy
University of Wisconsin
777 Highland Avenue
Madison, WI 53705 (USA)
[**] We would like to thank Dr. Kara Herlihy, Dr. Ralph Mazitschek, Dr.
Carlos Tassa, Jason Fuller, Dr. Steve Haggarty, Dr. Jianping Cui, Dr.
Letian Kuai, Dr. Marvin Yu, and Dr. Philip Yeske (Fluorous
Technologies) for reagents or comments. Work described herein has
been funded in whole or in part with Federal fundsfrom the
National Cancer Institute’s Initiative for Chemical Genetics,
National Institutes of Health, under contract no. N01-CO-12400.
Supporting information for thisarticle isavailable on the WWW
under http://www.angewandte.org or from the author.
Figure 1. Experimental approach to validating the use of fluorous-based
SMMsfor HDAC inhibitor discovery.
7960
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7960 –7964

Angewandte
Chemie
methods.[32,33] Finally, SMM binding
data, biochemical activity data, and
SPR data were compared to assess the
accuracy of fluorous microarrays in
identifying HDAC inhibitors.
Microarrays were printed with a set
of 20 fluorous-tagged molecules antici-
pated to be a mix of active and inactive
inhibitors (Scheme 1). Compounds 1F
to 3F are fluorous-tagged SAHA ana-
logues that serve as controls. The other
17 compounds are part of a collection
of candidate HDAC inhibitors with
varied linkers, metal chelators, and
affinities.[34]
Dimethylsulfoxide
(DMSO) and a fluorous-tagged com-
pound known to bind FKBP12 were
printed as negative controls.[21] We
probed the arrays with purified His-
tag fusions of HDAC2, HDAC3/
NCoR2 peptide complex (HDAC3/
NCoR2), and HDAC8 (we have deter-
mined that we are able to assess the
biochemical activity of these zinc-de-
pendent enzymes accurately). Arrays
were then incubated with an Alexa-
647-labeled anti-His antibody to detect
HDAC binding.
Fluorescence imaging revealed
nearly identical profiles for HDAC2
and HDAC3/NCoR2, whereas HDAC8
displayed
significant
differences
(Figure 2). Fluorescence intensity at
635 nm was measured for each printed
compound feature and averaged over at
least 30 replicates. Compounds display-
ing greater than twofold signal above
background (established with DMSO
controls) were classified as positives
(Figure 2). Compound 1F, a fluorous
SAHA analogue, displayed almost 10-
fold signal over background with
HDAC3/NCoR2 and 12-fold over back-
ground with HDAC2. The low-potency
free acid and methyl ester analogues of
SAHA (2F and 3F) showed signifi-
cantly lower signal in these profiles.
Eight other compounds in these two
profiles also displayed fluorescence
above the twofold threshold. Free
SAHA was also used in a competition
assay with HDAC3/NCoR2, which
markedly changed the array profile
(see the Supporting Information).
SAHA is known to be a weak inhibitor
of HDAC8, correlating with the
observed weak signal of 1F in the
profile. 11F is among the three com-
Scheme 1. Small molecules tested on microarrays, in biochemical activity assays, and with SPR assays.
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7960 –7964
www.angewandte.org
7961

Communications
pounds that showed significant signal over background in the
HDAC8 profile.
Non-fluorous analogues of each compound (1H–20H)
were then assessed for enzymatic inhibition by an established
biochemical activity assay (Figure 3).[35] Ten compounds for
HDAC2 and nine compounds for HDAC3/NCoR2 demon-
Figure 3. Biochemical activity assay data for HDAC2, HDAC3/NCoR2
complex, and HDAC8. Elementshighlighted in red mark compounds
classified as positives on SMMs.
strated 10% inhibition or greater at 333 nm. As anticipated,
only compounds with metal-chelating elements, such as
hydroxamates and ortho-hydroxy anilides proved to be
effective inhibitors of these enzymes. Results from biochem-
ical activity assays and SMM assays were congruent, with
eight of ten inhibitors (80%) for HDAC2 and eight of nine
(89%) for HDAC3/NCoR2 also classifying as positives on the
SMMs. Compound 16H, which demonstrated no inhibitory
activity at 333 nm but whose analogue 16F was classified as a
positive, showed considerable inhibitory activity at 3.33 mm
(data not shown). For HDAC8, only four compounds showed
greater than 20% inhibition, with six weaker inhibitors falling
between 10–20% inhibition. Unexpectedly, three of these
weaker inhibitors were methyl ester analogues. Fifty percent
of the strongest inhibitors (2/4) of HDAC8 also classified as
positives on the SMMs, showing good agreement between the
data sets.
Figure 2. SMM data for HDAC2, HDAC3/NCoR2, and HDAC8: a) The
arrayswere probed with protein followed by an Alexa-647-labeled anti-
pentaHis antibody. b) The histograms represent fold signal intensities
with respect to the background signal, established from features
containing DMSO only (D in array key). Valuesare averagesof at least
30 replicates. Red bars indicate intensities that are more than twofold
greater than the background signal and classify as positives.
7962 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7960 –7964

Angewandte
Chemie
For a few compounds, the data derived from microarray
and biochemical activity assays for each of the HDACs did
not correlate well. To account for these differences, SPR
experiments were conducted with HDAC3/NCoR2 to exam-
ine the thermodynamic and kinetic binding behavior of these
compounds (Figure 4). SAHA was first rigorously character-
ized with HDAC3/NCoR2 to establish that the enzyme was
competent for binding while displayed on the surface
(Figure 4a,b). The empirically determined dissociation con-
stant of 22 nm correlates with previously published IC50
values, providing confidence in the assay.[6]
The remaining 19 compounds were then evaluated at
three different concentrations to rank their affinities and
binding kinetics. The non-fluorous analogues corresponding
to positives in the SMM experiments displayed significant
binding in an SPR-based ranking assay (Figure 4c).[32–33]
Compound 8H displayed 50% enzymatic inhibition yet its
fluorous analogue did not classify as a positive. We note that
8H also had the fastest relative dissociation rate constant of
the compounds tested (Figure 4d). Discrepancies between
the different data sets may be explained by an inability of the
microarrays to identify enzyme binders with relatively fast
dissociation rates.
Previous studies have demonstrated that small-molecule
microarrays can be used effectively with whole-cell lysates.[21]
To test if fluorous microarrays can be used to detect native
HDACs, arrays were incubated with whole-cell lysates from
293-MSR cells. Since HDAC3 is present in 293-MSR cells,[36]
arrays were probed with mouse monoclonal anti-HDAC3
antibody mixed with Alexa-647-labeled goat anti-mouse
antibody (Figure 5). Six of the seven positives on these
Figure 4. Compounds 1H-20H were tested for binding to HDAC3/
NCoR2 by using SPR methods: a) SAHA was characterized (n=3) by
measuring binding in a dilution series (3 nm to 729 nm). Thermody-
namic and kinetic analyses of these curves yielded binding constants:
k
on =4.9”105 mÀ1 sÀ1, koff =9.18”10À3 sÀ1, KD =22 nm. b) Plot of con-
centration versus response from SAHA dilution series used to calcu-
late the equilibrium dissociation constant. c) Plot showing compound
affinities at three concentrations. Red asterisks indicate compounds
scored as positives in the SMM experiments. d) Plot of kon versus koff
for compounds with measurable kinetics from the SPR ranking assay.
Figure 5. Small-molecule microarray data for 293-MSR cell lysate:
a) Imagesof arraystreated with lysate and purified HDAC3/NCoR2.
b) Histogram of fold signal intensities with respect to the background
signal for lysate-treated arrays. Red bars indicate positives and
asterisks indicate compounds that were positives with purified
HDAC3/NCoR2.
Angew. Chem. Int. Ed. 2007, 46, 7960 –7964
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7963

Communications
[15] A. D. Tran, T. P. Marmo, A. A. Salam, S. Che, E. Finkelstein, R.
Kabarriti, H. S. Xenias, R. Mazitschek, C. Hubbert, Y. Kawa-
guchi, M. P. Sheetz, T. P. Yao, J. C. Bulinski, J. Cell Sci. 2007, 120,
1469.
[16] V. S. Santander, C. G. Bisig, S. A. Purro, C. H. Casale, C. A.
Arce, H. S. Barra, Mol. Cell. Biochem. 2006, 291, 167.
[17] T. Hideshima, J. E. Bradner, J. Wong, D. Chauhan, P. Richard-
son, S. L. Schreiber, K. C. Anderson, Proc. Natl. Acad. Sci. USA
2005, 102, 8567.
[18] S. J. Haggarty, K. M. Koeller, J. C. Wong, C. M. Grozinger, S. L.
Schreiber, Proc. Natl. Acad. Sci. USA 2003, 100, 4389.
[19] J. R. Cabrero, J. M. Serrador, O. Barreiro, M. Mittelbrunn, S.
Naranjo-Suarez, N. Martin-Cofreces, M. Vicente-Manzanares,
R. Mazitschek, J. E. Bradner, J. Avila, A. Valenzuela-Fernandez,
F. Sanchez-Madrid, Mol. Biol. Cell 2006, 17, 3435.
[20] J. L. Duffner, P. A. Clemons, A. N. Koehler, Curr. Opin. Chem.
Biol. 2007, 11, 74.
arrays also classified as positives with purified HDAC3/
NCoR2, showing good agreement.
In conclusion, there is a strong correlation between
1) small molecules that bind HDACs identified from fluo-
rous-based SMMs, 2) inhibitors identified with biochemical
activity assays, and 3) binders identified from SPR assays.
Fluorous-based SMMs therefore offer a viable method for
discovering novel HDAC inhibitors in the future. Profiles
generated from these arrays against different HDAC homo-
logs may aid in the discovery of selective inhibitors, which is a
particularly important challenge in modern chromatin
research.
Received: July 17, 2007
Published online: September 17, 2007
[21] J. E. Bradner, O. M. McPherson, R. Mazitschek, D. Barnes-
Seeman, J. P. Shen, J. Dhaliwal, K. E. Stevenson, J. L. Duffner,
S. B. Park, D. S. Neuberg, P. Nghiem, S. L. Schreiber, A. N.
Koehler, Chem. Biol. 2006, 13, 493.
[22] N. Kanoh, M. Kyo, K. Inamori, A. Ando, A. Asami, A. Nakao,
H. Osada, Anal. Chem. 2006, 78, 2226.
[23] M. Köhn, R. Wacker, C. Peters, H. Schroder, L. Soulere, R.
Breinbauer, C. M. Niemeyer, H. Waldmann, Angew. Chem.
2003, 115, 6010; Angew. Chem. Int. Ed. 2003, 42, 5830.
[24] M. R. Lee, I. Shin, Angew. Chem. 2005, 117, 3317 – 2941; Angew.
Chem. Int. Ed. 2005, 44, 2881.
[25] M. Uttamchandani, D. P. Walsh, S. Q. Yao, Y. T. Chang, Curr.
Opin. Chem. Biol. 2005, 9, 4.
[26] K. S. Ko, F. A. Jaipuri, N. L. Pohl, J. Am. Chem. Soc. 2005, 127,
13162.
[27] D. P. Curran, Q. Zhang, C. Richard, H. Lu, V. Gudipati, C. S.
Wilcox, J. Am. Chem. Soc. 2006, 128, 9561.
[28] W. Zhang, Curr. Opin. Drug Discov. Devel. 2004, 7, 784.
[29] W. Zhang, Comb. Chem. High Throughput Screening 2007, 10,
219.
[30] W. Zhang, Y. Lu, J. Comb. Chem. 2006, 8, 890.
[31] W. Zhang, Y. Lu, T. Nagashima, J. Comb. Chem. 2005, 7, 893.
[32] D. G. Myszka, Anal. Biochem. 2004, 329, 316.
[33] W. Huber, J. Mol. Recognit. 2005, 18, 273.
[34] S. M. Sternson, J. C. Wong, C. M. Grozinger, S. L. Schreiber,
Org. Lett. 2001, 3, 4239.
[35] B. E. Schultz, S. Misialek, J. Wu, J. Tang, M. T. Conn, R.
Tahilramani, L. Wong, Biochemistry 2004, 43, 11083.
[36] S. Emiliani, W. Fischle, C. Van Lint, Y. Al-Abed, E. Verdin, Proc.
Natl. Acad. Sci. USA 1998, 95, 2795.
[37] R. L. Nicholson, M. L. Ladlow, D. R. Spring, Chem. Commun.
2007, DOI:10.1039/b712906h.
Keywords: chromatin · fluoroustags·
high-throughput screening · microarrays
.
[1] T. Kouzarides, Cell 2007, 128, 693.
[2] B. E. Bernstein, A. Meissner, E. S. Lander, Cell 2007, 128, 669.
[3] A. J. de Ruijter, A. H. van Gennip, H. N. Caron, S. Kemp, A. B.
van Kuilenburg, Biochem. J. 2003, 370, 737.
[4] J. Taunton, C. A. Hassig, S. L. Schreiber, Science 1996, 272, 408.
[5] P. A. Jones, S. B. Baylin, Cell 2007, 128, 683.
[6] A. Vannini, C. Volpari, G. Filocamo, E. C. Casavola, M.
Brunetti, D. Renzoni, P. Chakravarty, C. Paolini, R. De Fran-
cesco, P. Gallinari, C. Steinkuhler, S. Di Marco, Proc. Natl. Acad.
Sci. USA 2004, 101, 15064.
[7] C. M. Grozinger, S. L. Schreiber, Chem. Biol. 2002, 9, 3.
[8] E. Hockly, V. M. Richon, B. Woodman, D. L. Smith, X. Zhou, E.
Rosa, K. Sathasivam, S. Ghazi-Noori, A. Mahal, P. A. Lowden,
J. S. Steffan, J. L. Marsh, L. M. Thompson, C. M. Lewis, P. A.
Marks, G. P. Bates, Proc. Natl. Acad. Sci. USA 2003, 100, 2041.
[9] P. A. Marks, X. Jiang, Cell Cycle 2005, 4, 549.
[10] P. A. Marks, V. M. Richon, W. K. Kelly, J. H. Chiao, T. Miller,
Novartis Found. Symp. 2004, 259, 269.
[11] W. K. Kelly, V. M. Richon, O. OꢀConnor, T. Curley, B. MacGre-
gor-Curtelli, W. Tong, M. Klang, L. Schwartz, S. Richardson, E.
Rosa, M. Drobnjak, C. Cordon-Cordo, J. H. Chiao, R. Rifkind,
P. A. Marks, H. Scher, Clin. Cancer Res. 2003, 9, 3578.
[12] L. M. Krug, T. Curley, L. Schwartz, S. Richardson, P. Marks, J.
Chiao, W. K. Kelly, Clin. Lung Cancer 2006, 7, 257.
[13] G. Garcia-Manero, J. P. Issa, Cancer Invest. 2005, 23, 635.
[14] J. C. Wong, R. Hong, S. L. Schreiber, J. Am. Chem. Soc. 2003,
125, 5586.
7964 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 7960 –7964